Abstract:

Matched Doppler processing (MDP) is a new technique that estimates the range and speed of a near-field narrow-band sound source as it moves past an underwater acoustic sensor by coherently matching the received signal against replicas. Replica vectors are generated according to a simple propagation model, which accounts for the effect of the Doppler phenomena on the signal phase. The data to be matched are the vector of complex outputs of Fourier transforms of contiguous blocks of hydrophone data and the matching is by means of a cost or correlation function, which processes the signal phase coherently. The main contribution reported is the extension of MDP to an array of receivers for localization of a source's azimuth and heading as well as range and speed. Two methods are considered: The first method performs single receiver MDP on each sensor independently, then sums the cost values. The second method processes the signal phase coherently across space and also time. Extensive simulations compare the performance of both methods for an array of varying apertures and the number of receivers. Further simulation indicates that the second method affords sidelobe suppression, which improves performance when multiple signals are present. Application to tracking multiple sources is suggested and discussed.